In 1990, we are much less certain that we understand ARDS than we were in 1982, and we have yet to identify specific therapy. It is tempting to conclude that we have made no progress, but this conclusion would be unwarranted. In those 8 years, important advances have been made. The complement hypothesis has survived, with significant modifications. Recognition of the importance of infection in clinical outcome and of endotoxin in augmentation of neutrophil-mediated injury has evolved in concert and meshes well. A new class of peptide mediators, cytokines, has assumed a central role. Lipid mediators now appear as modulators of cytokine-induced effects by priming, amplification, and regulation of gene expression rather than as unifactorial "causes" of the physiologic manifestations of ARDS. These interdigitating mechanisms have been recognized as pansystemic, resulting in overt multiple organ dysfunction and ultimately in death if amplification mechanisms go unchecked. Technologies in molecular genetics, generally unknown to the pulmonary community in 1982, have had a significant impact. Recombinant cDNA technology has permitted identification of the existence, structure, and functions of novel cytokines; made them available in sufficient quantity for detailed study; and prompted interest in the regulation of gene expression in the evolution and resolution of inflammation. Proteins modified by genetic engineering, as well as monoclonal antibodies and receptor antagonists for specific cytokines, are promising future approaches to therapy. At present, the complexity of the redundant networks by which inflammation is regulated seems bewildering in relation to ARDS. Bewildering or not, the age of the "mediator" of ARDS, and of the corresponding therapeutic "magic bullet," is over. The complexity of the system of regulatory checks and balances must be addressed at the molecular level.